The goals of this project are to evaluate neurobehavioral outcomes of children who sustain traumatic brain injury (TBI). The hypothesis is that the nature and extent of injury to the frontal lobes account for significant variability in outcome beyond age and injury severity. Based on MRI findings, the relationship of focal brain lesions to cognitive sequelae, particularly functions such as working memory, are investigated. The Retrospective Study (Study I) incorporates a cross-sectional design to investigate cognitive and psychosocial outcome in children and adolescents who were injured at least 3 years prior to the assessment. Although the selection criteria specify a minimum age of 5 years to complete the outcome measures, the Retrospective Study provides an opportunity to include children who were injured as infants and to investigate long term outcome. The Retrospective Study enables us to replicate reports suggesting that early prefrontal injury has a particularly devastating impact on psychosocial development. The Prospective Study (Study II) utilizes a longitudinal design to investigate developmental changes in cognition and behavior in relation to the age at injury, the severity of CHI, and the volume of frontal vs extrafrontal brain lesions. Several studies were completed that utilized CRC support. The first (Levin et al., Cortex, in press) addressed recovery of semantic memory after head injury. Semantic memory refers to the organized body of general knowledge which people possess about words, concepts, facts, and the rules for manipulating these symbols and concepts. We analyzed several measures of semantic memory in relation to the severity of injury, age at injury, and the presence and volume of focal brain lesions in cross-sectional (n=190) and longitudinal (n=94) studies. The effects of TBI severity were confirmed on most measures of semantic memory in the longitudinal study as was a pattern of recovery from 3 to 12 months postinjury. Using a series of multiple regressions, we investigated whether entering the volume of focal brain lesions incremented the severity of injury and age at injury in predicting performance on tests of semantic memory. The volume of left extrafrontal lesions most consistently incremented the prediction of semantic memory on the 12 month assessment (the most common sites of extrafrontal lesions were in the temporal and parietal regions). A similar pattern was found for the multiple regressions at 3 months. A second study (Fletcher et al., J Child Neurol, in press) investigated the behavioral adjustment and adaptive functioning of 138 children with mild to severe TBI who ranged in age from 6 to 16 years at the time of the assessment (divided into age groups of 6-10, 11-16 years) and were at least 3 months postinjury (the mean postinjury interval exceeded one year). The goal of the study was to investigate the relationship of two measures of psychosocial outcome to indices of TBI severity and MRI findings. Each child's caregiver completed the Personality Inventory for Children-Revised and 77 of the children were also studied using the Vineland Adaptive Behavior Scale. Of the four broad-band factor scales of the Personality Inventory for Children (i.e., Undisciplined, Poor Self-Control; Social Incompetence; Internalization/Somatic Symptoms; and Cognitive Development), only Cognitive Development was sensitive to the severity of CHI (p<.001). Analysis of the Vineland Adaptive Behavior Scale disclosed significant effects of TBI severity on the Communication and Socialization domains, whereas the scores of the severely injured patients did not differ from the results obtained for the mild/moderate group on the Daily Living domain. A significant age effect was found for the Daily Living domain as younger children had lower scores than the 11-16 year group. The third study (Levin et al., Dev Neuropsychol, in press) demonstrated that executive function tasks do not measure unitary dimensions of cognition, but are multidimensional. The results of the semantic memory study indicate that the deficits following severe head injury are not limited to learning material such as word lists, but may also extend to the capacity to acquire general knowledge. Our study of semantic memory is potentially relevant to the educational outcomes of pediatric TBI because it relates to the capacity of children for acquiring general knowledge similar to the learning which occurs in the classroom situation. In contrast to our previous analysis of measures of executive function, the volume of extrafrontal lesions (primarily temporoparietal) incremented the prediction of semantic memory more consistently than the volume of regions of abnormal signal in the frontal lobes. Our recent study of psychosocial outcome confirms the sensitivity of the Vineland Scale to severity of TBI and implicates the usefulness of the Vineland as a global outcome measure for clinical trials involving pediatric head injury. Using age-corrected percentile or standard scores, several domains of the Vineland are clearly sensitive to TBI severity and permit direct comparison of outcomes in children who vary in age at injury. Finally, executive function tasks are multi-factorial, supporting the need for component analyses of task performance. Our plans include continued accrual of patients in the Retrospective and Prospective Studies. Our goals for publication during the 1995-1996 period include a paper concerning serial changes in neuropathology on repeat MRI. We also plan to analyze longitudinal data for the discourse component of the project, including an analysis of the relationship of TBI severity, age at injury, and focal brain lesions to recovery.